- in-house algorithm
- short & long reads
- mixing of paired reads (variable insert size & orientation
- quality trimming!
- hybrid data (454 & Illumina)
- the more data the bigger the k-mer size
- they say 'optimal parameter selection sometimes requires experimentation'
- -g option --> use certain reads only in later stages (gap closing, bubble resolving, scaffolding)
-
Comparison of assemblies from different tools at the end
-
?? Downsampling of input data -> use max coverage
-
TODO - test with gage data
-
TODO - testing framework